Completed support for tape density and add usage of at() in a vector access to prevent segfault in unit test.
Removed some throw() specifications which led to segfaults.

Implemented cta::ArchiveJob::complete(), and added the members that should be set to have the necessary
information available at complete() time.
Updated the calls in the archive_and_retrieve_new_file unit test.

Fixed the MigrationReportPackerTest and changed the MigrationReportPacker to be more similar to the RecallReportPacker

The SchedulerDB provides a TapeMountDecisionInfo object, which contains all that the schduler needs to decide what
tape mount to start. The tape mount decision info then creates an archive or a retrieve mount with 2 separate methods.
The schedulerDB does not have a bare TapeMount class any more. On the scheduler itself will.

The bottom line is the Mount and Job classes shall be passed around the
tapeserverd daemon for two purposes, to carry all the information about
the file copy to be transfered and to provide the notification interface
to the scheduler on which the tapeserverd daemon needs to call methods
to indicate the ends of transfers and mounts.

Improved the readbility of the Scheduler code by replacing dynamic casts with the addition of a simple getMountType() method to the TapeMount, ArchiveMount and RetrieveMount classes

Moved mock scheduler DB in a separate directory to reduce the number of files per directory.

Moved all of the classes in middletier/common to the scheduler directory because most of them are specific to the scheduler.  Please feel free to move those that are no so specific elsewhere.

middletier, itself separated in interface, SQLite, objectstore and
shared tests.
Moved all utilities (exceptions, threading...) to a shared utility
directory.
Created a single, shared unit test from all the scattered ones.

License comment is no longer filtered out by doxygen when generating web pages of the individual source files.